Sourcegraph Cody
Sourcegraph Cody is an AI coding assistant that grounds chat, completions, and agent workflows in your entire multi-repo codebase.
Cody's bet is that context beats cleverness: grounding completions, Deep Search, and coding agents in real symbols across hundreds of repositories is hard to replicate with a lighter assistant. The 2026 releases sharpen that bet — Code Finder is GA as an MCP tool that hands agents exact file and line ranges, Deep Search writes downloadable CSV/JSON/SVG reports, and Agentic Batch Changes is now priced only on changesets that actually merge. The catch is the entry price: Enterprise starts at $16K on a minimum annual contract, which makes this an org purchase, not a weekend experiment. Pick it when codebase scale and agent context are the bottleneck; skip it when one repo and a $10–20/month
Verified 8d ago · liveness 87/100 · cite: rightaichoice.com/tools/sourcegraph-cody
- Enterprise engineering orgs navigating large, multi-repo codebases
- Platform and migration teams changing code across hundreds of repositories
- Teams running Claude Code, Codex, Cursor, or Amp agents that need structured codebase context via MCP
- Security teams triaging findings across many repos with admin controls
- Solo developers and small single-repo projects that a lightweight per-seat assistant already covers
- Buyers whose primary comparison is a $10–20/month per-seat coding assistant
- Teams that require fully offline operation or zero telemetry
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Sourcegraph Cody if you work in a single repository and want a per-seat AI assistant you can expense, since Cody is sold as a Sourcegraph Enterprise contract starting at $16K on a minimum annual term.
AI features are metered through credits, so a team burning heavily on chat and Deep Search can exhaust its included per-user allocation before the contract term ends
Cody is priced for enterprise engineering organizations: Enterprise starts at $16K on a minimum annual contract, with included AI credits, org-wide pooling, self-hosted or single-tenant cloud deployment, and 24×5 support. That puts it well above per-seat assistants that run roughly $10–20 per user per month, and closer to other enterprise code-intelligence and platform contracts. It fits orgs where codebase scale, agent context via MCP, and codebase-wide change execution justify a platform
In short
Sourcegraph Cody — Sourcegraph Cody is an AI coding assistant that grounds chat, completions, and agent workflows in your entire multi-repo codebase. Best for Enterprise engineering orgs navigating large, multi-repo codebases, Platform and migration teams changing code across hundreds of repositories, Teams running Claude Code, Codex, Cursor, or Amp agents that need structured codebase context via MCP. Plans from $16/yr.
What's new in Sourcegraph Cody
Checked 8 days agoAcross the latest 5 updates: 2 feature updates, 1 launch, 1 pricing change and 1 changelog entry.
Talk to your agents
You can now dictate prompts in Deep Search and Agentic Batch Changes, so typing speed stops being the limit on expressing what you want the agent to do.
Coding agents usually can't price on outcomes. Ours can.
Agentic Batch Changes is now priced on outcomes — customers pay only for the changesets that merge, rather than for agent activity that goes nowhere.
Introducing Agentic Batch Changes: the frontier agent for code change at scale
Agentic Batch Changes launches, letting one engineer run codebase-wide changes across hundreds or thousands of repositories — aimed at migrations, upgrades, and security fixes that keep getting deferred.
Code Finder adds revision-specific search; GitHub Enterprise Server 3.3 support dropped
Code Finder can now search a specified branch, tag, or commit for investigating regressions at exact commits or non-default release branches, and the minimum supported GitHub Enterprise Server version is now 3.4.
Code Finder is now generally available
Code Finder, an agent-optimized MCP code search tool, exits beta and is generally available via the Sourcegraph MCP server, returning a concise summary with links and line ranges to coding agents.
What people actually say about Sourcegraph Cody — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
14 mentions across 1 source (Hacker News) · researched Jul 31, 2026.
Average across the 1 source that answered — each source counts once, not each post.
- +Deep codebase context across multiple repos is a genuine edge.
- +Integrates with VS Code, JetBrains, Visual Studio, and web app.
- +Smart hover summaries provide instant symbol documentation.
- +MCP server and Search API enable powerful agentic workflows.
- +Supports multiple LLMs including Claude Sonnet 4.6.
- −Costs can spiral with heavy LLM usage, burning credits fast.
- −Context building is opaque and lossy, frustrating at times.
- −Setup is more complex than Copilot or Cursor for new users.
- −Not ideal for small projects or solo devs—overkill.
- −Occasional accuracy issues if context is incomplete.
- • LLM API usage can drive costs beyond subscription? (Unclear from data).
- • Enterprise tier may require annual contracts.
Viability Score
How well maintained and how widely used is Sourcegraph Cody? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Developer chat with context from your open file and repository by default
- @-mention context for specific files, symbols, remote repositories, and non-code artifacts
- Code completions and code edits
- Auto-edit suggestions based on cursor movement and recent typing
- Premade and customizable Prompts you can save and share with your team
- Context Filters to exclude selected repositories from chat and autocomplete
- Debugging assistance optimized to identify and fix errors
- Code Finder MCP tool (generally available) returning concise summaries with links and line ranges
- Code Finder revision search across a specified branch, tag, or commit (August 31, 2026)
- Agentic Batch Changes using AI agents to make, review, and ship changes across repositories
- Deep Search AI agent that explores your codebase and aggregates larger result sets
- Deep Search downloadable CSV, JSON, and SVG reports, readable back in follow-up investigations
- Deep Search customizable guidance (AGENTS.md-style) and task-specific skills
- Guided diff tour (Beta) grouping related files into a suggested reading path
- Repository Overview showing repository status at a glance
About Sourcegraph Cody
Sourcegraph Cody is an AI coding assistant sold on the Sourcegraph Enterprise platform. It combines developer chat, code completions, code edits, auto-edit suggestions, and shareable Prompts with Sourcegraph's Search API, so answers draw on real APIs, symbols, and usage patterns from local and remote repositories instead of generic training patterns. Cody runs in VS Code, JetBrains, Visual Studio (experimental), the Sourcegraph web app, and from the terminal via Cody CLI, and connects to code hosts including GitHub and GitLab. Chat carries your open file and repository context by default, and you can @-mention specific files, symbols, remote repositories, or non-code artifacts. Auto-edit starts proposing changes once you have made at least one character edit, based on cursor position and recent typing; Context Filters let admins and users exclude selected repositories from chat and autocomplete. Sourcegraph says it collects your prompts and responses to provide the service but does not use your data to train models. Pricing is Enterprise-only: starting at $16K on a minimum annual contract, with included AI credits, org-wide credit pooling, no monthly credit expiry, volume credit buckets, plus single-tenant cloud or self-hosted deployment and admin controls. The platform around Cody has moved fast in 2026 — Agentic Batch Changes, Deep Search with downloadable CSV/JSON/SVG reports, and Code Finder (an MCP code-search tool now generally available) are the current center of gravity, and Agentic Batch Changes is now priced on outcomes, with customers paying only for changesets that merge.
Behind the Verdict
Cody is best understood as the assistant layer on top of Sourcegraph's code intelligence, not as a standalone autocomplete toy. Chat defaults to your open file and repository context, and @-mentions pull in specific files, symbols, remote repositories, or non-code artifacts — so an answer about an internal API can actually cite how that API is used elsewhere in your org rather than inventing a plausible signature. Strengths. The context pipeline is the differentiator, and it keeps getting sharper. Code Finder, now generally available via the Sourcegraph MCP server, is an agent-optimized code search tool that returns a concise summary with links and exact line ranges, and as of the August 31, 2026 update it can search a specified branch, tag, or commit — useful for investigating regressions at an exact commit or digging into a non-default release branch. Deep Search now cross-references and aggregates larger result sets and writes findings to downloadable CSV, JSON, and SVG files, so codebase-wide inventories and migration audits no longer have to fit in a chat answer; it can also read generated CSV/JSON back in follow-up runs. You can give Deep Search durable AGENTS.md-style guidance plus task-specific skills, configured globally, per-org, or personally, and admins can define an exclusion list so filtered repos and files are invisible to it. On the review side, Guided diff tour (Beta) groups related files into sections, orders them into a suggested reading path, and adds short explanations. Agentic Batch Changes pushes all of this into codebase-wide change execution — and is now priced on outcomes, with customers paying only for changesets that merge. Constraints worth naming. Everything above is Enterprise-gated: pricing starts at $16K on a minimum annual contract, with AI features metered through credits that include per-user allocation and org-wide pooling. Operations teams should note the deployment surface (single-tenant cloud or self-hosted) and the admin configuration Cody needs — code host connections and access setup come before cross-repo context works. Sourcegraph also enforces geographic access restrictions, which it reinforced in an August 2026 update, and it dropped support for GitHub Enterprise Server 3.3 and below. On privacy, Sourcegraph states it collects prompts and responses to provide the service but does not train models on your data. Where it fits. Multi-repo engineering organizations where the bottleneck is understanding and changing code at scale: platform teams running migrations, security teams triaging findings across many repositories, and anyone running Claude Code, Codex, Cursor, or Amp agents that need structured codebase context through MCP. Where it does not: single-repo teams, solo developers, or buyers whose primary comparison is a per-seat monthly subscription, and teams that need fully offline operation or zero telemetry.
Researching Sourcegraph Cody? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Sourcegraph Cody actually fits — and what changes day-one when you adopt it.
You need to update a deprecated internal HTTP client across 400 repositories before a deadline. You open Agentic Batch Changes, describe the change, and let the agent run across repositories while you review proposed diffs — Guided diff tour groups related files and orders them into a reading path for the large ones.
Outcome: The migration ships as reviewable per-repo changesets instead of a quarter-long manual sweep, and because Agentic Batch Changes is priced on outcomes you pay for the changesets that actually merge.
A scanner flags a vulnerable pattern in dozens of services. You run Deep Search with your organization's AGENTS.md-style guidance, asking for a codebase-wide inventory; Deep Search aggregates the results and attaches a CSV checklist you can filter and enrich in a follow-up run.
Outcome: You get a structured, exportable inventory of every affected location instead of a chat answer truncated by result size, and admin-defined context filters keep restricted repos out of the search entirely.
You install the Cody extension in VS Code, ask chat how a service authenticates, and @-mention the symbol plus a remote repository. When your coding agent needs an exact implementation, you point it at Code Finder, which returns a concise summary with links and line ranges.
Outcome: You get answers grounded in the real symbols and usage patterns in your org's code rather than plausible-sounding generated APIs, and your agent gets focused context instead of a token-heavy file dump.
Use Cases
- Onboarding to a large monorepo by asking Cody about code structure, symbols, and where an API is actually used
- Debugging a runtime error by using Cody to analyze stack traces and the surrounding code
- Refactoring a function across multiple files with Cody's multi-repo context
- Running Agentic Batch Changes to update an API or dependency across hundreds of repositories
- Producing a codebase-wide migration or audit report with Deep Search and exporting it as CSV, JSON, or SVG
- Investigating a regression at an exact commit or non-default release branch with Code Finder revision search
- Reviewing a large diff faster using Guided diff tour's grouped, ordered reading path
- Giving Claude Code, Codex, Cursor, or Amp agents structured codebase context through the Sourcegraph MCP server
Models Under the Hood
as of 2026-09-15
Limitations
- Cody is supported on Sourcegraph Enterprise, with pricing starting at $16K on a minimum annual contract — so it is an organizational purchase rather than an individual subscription.
- AI features are metered through credits: the plan includes a per-user credit allocation with org-wide pooling, and you should model your usage against the Enterprise Portal's credit accounting before committing.
- You need to connect your code hosts and configure repository access before context across codebases works.
- GitHub Enterprise Server 3.3 and below are no longer supported; the minimum is 3.4.
- Sourcegraph also enforces geographic access restrictions, reinforced by an August 2026 update, and says it collects prompts and responses to provide the service, though it does not train models on your data.
as of 2026-09-30
Verification history
We have re-verified Sourcegraph Cody 21 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
Showing the 6 most recent of 21 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Sourcegraph Cody tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Enterprise
Starting at $16K/yr (minimum annual contract)
Ideal for
Engineering organizations with large, multi-repo codebases that need centralized code intelligence, agent context via MCP, and admin/security controls
What this tier adds
Starting tier: includes AI credits with org-wide pooling, Deep Search, Agentic Batch Changes, full MCP server/API/CLI access, single-tenant cloud or self-hosted deployment, and 24×5 support, on a minimum annual contract starting at $16K
Where the pricing makes sense
The company stage and team size where Sourcegraph Cody's pricing actually pencils out — and where peers do it cheaper.
Cody is priced for enterprise engineering organizations: Enterprise starts at $16K on a minimum annual contract, with included AI credits, org-wide pooling, self-hosted or single-tenant cloud deployment, and 24×5 support. That puts it well above per-seat assistants that run roughly $10–20 per user per month, and closer to other enterprise code-intelligence and platform contracts. It fits orgs where codebase scale, agent context via MCP, and codebase-wide change execution justify a platform
Setup time & first value
How long it actually takes to get something useful out of Sourcegraph Cody — broken out by persona, not the marketing-page minute.
For an individual developer inside an already-configured Sourcegraph Enterprise instance: minutes — install the VS Code, JetBrains, or Visual Studio extension, or use the web app or Cody CLI, and chat works with your open file and repository immediately. For the organization: days to weeks. An admin must connect code hosts such as GitHub or GitLab, configure repository access and Context Filters,
Switching to or from Sourcegraph Cody
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From GitHub Copilot or a per-seat chat assistant: keep your editor, install the Cody extension, and route questions that need cross-repo context to Cody chat and Deep Search
- →From manual code search and grep: move repeated cross-repo lookups to Code Search and Deep Search, then export the larger investigations as CSV, JSON, or SVG reports
- →From scripted multi-repo edits: replace bespoke scripts with Agentic Batch Changes and review the resulting changesets in-platform
- →From a general-purpose coding agent: point it at the Sourcegraph MCP server so it can call Code Finder for exact file and line ranges instead of reading whole directories
- ↗To a per-seat assistant like a $10–20/month coding subscription: expect to lose cross-repo grounding, Deep Search aggregation, and the MCP code graph
- ↗To a general-purpose agent plus local repo access: keep single-repo work moving, but you give up org-wide symbol and usage context
- ↗To manual code review tooling: Guided diff tour and Agentic Batch Changes are the pieces you would be replacing by hand
Integrations
Resources & Guides
Tutorials & Learning
Sourcegraph Cody Tutorial for Beginners | Coding Assistant Setup & Demo
BizGuide
Sourcegraph Codyレビュー：最強のAIコーディングアシスタントを徹底検証
TWiz
Master Any Codebase with Sourcegraph Cody: A Developer’s Guide
PixelCraft
YouTube returned 6 videos for “Sourcegraph Cody”, and we withheld 1: 1 did not mention Sourcegraph Cody. Showing the 5 we can prove are about Sourcegraph Cody.
Official links
Tools that pair well with Sourcegraph Cody
Common stack mates teams adopt alongside Sourcegraph Cody, with the specific reason each pairing earns its keep.
Claude
Claude is Anthropic's AI assistant for long-document analysis, coding, and agentic work in one chat-and-Cowork surface.
JetBrains AI
JetBrains AI puts agent-driven coding assistance, multi-agent choice, and enterprise governance inside the JetBrains IDE family
Chatbox
Chatbox AI is a multi-model chat client with a local knowledge base, real-time web search, and a desktop Work Mode agent
Featured Head-to-Head Comparisons
Sourcegraph Cody vs Windsurf
Pick Sourcegraph Cody if your pain is codebase context—you work across many repos, need the AI grounded in your actual APIs, and want admin controls like RBAC. Pick Windsurf if your pain is agent sprawl—you juggle local and cloud agents and want one command center with shared context. If you’re a solo dev on a small project, Windsurf’s free unlimited SWE-1.7 and simpler setup beats Cody’s enterprise tilt.
Claude vs Sourcegraph Cody
Choose Sourcegraph Cody if your team lives in multi-repo codebases and needs deep integration with GitHub/GitLab, with agentic search and batch changes. Choose Claude if you need a versatile AI for heavy document analysis, long reports, and design-to-code, with the new Opus 5 offering near-flagship power at half cost. For pure coding, Cody’s code-specific context wins; for general analysis and coding, Claude’s breadth wins.
Cursor vs Sourcegraph Cody
If your team lives in large, multi-repo codebases and needs deep code understanding with enterprise controls, Sourcegraph Cody is the pragmatic pick. But if you want an AI that takes a feature from idea to deployment with minimal hand-holding, Cursor's agentic autonomy is unmatched — just be ready to pay $20/mo and accept cloud dependency.
Codeium vs Sourcegraph Cody
If you're an individual developer or small team who wants to start free and grow into agent workflows, Codeium is the obvious pick — unlimited Tab completions and inline edits on the Free plan, with a full IDE (Devin Desktop) and an Agent Command Center once you need to run fleets of agents. If you're an engineering org working across dozens of repos that needs answers grounded in real symbols and APIs, plus self-hosted deployment, Sourcegraph Cody's $16K/yr enterprise contract is the more appropriate tool. The two barely overlap in buyer: Codeium competes with Copilot and Cursor on price and breadth; Cody competes in the enterprise code-intelligence space. Don't compare them on sticker price alone — compare them on whether you need a per-seat assistant or an org-wide grounded codebase layer.
Alternatives to Sourcegraph Cody
View allClaude
Claude is Anthropic's AI assistant for long-document analysis, coding, and agentic work in one chat-and-Cowork surface.
JetBrains AI
JetBrains AI puts agent-driven coding assistance, multi-agent choice, and enterprise governance inside the JetBrains IDE family
Frequently Asked Questions
Categories
Topics
Used Sourcegraph Cody? Help shape our editorial sentiment research.